Data-source registry
Every source Helios is permitted to read is declared here before any code fetches from it, together with its licence, rate limit, and historical depth. The sources Helios cannot read are listed too, each with the reason. The gaps are published deliberately: a site with thin evidence might be a quiet project or a blocked source, and this is the only page that tells the two apart.
Geographic coverage
1 of 9 regions readHelios names more regions than it reads. A region listed as 'declared' is in scope and holds no data; only 'active' regions have connectors behind them. Site counts are the check on that claim.
| Region | Counties | Coverage | Sites |
|---|---|---|---|
| East Valley, Arizona The pilot region. Chosen because Maricopa County publishes parcel geometry, ownership and land-use classification openly, which makes it possible to check Helios against records anyone else can read. | Maricopa, Pinal | active | 13 |
| West Valley, Arizona Same county records as the pilot, so the marginal cost of covering it is a city list rather than a connector. | Maricopa | declared | 0 |
| Northern Virginia The largest concentration of data-centre capacity in the world. | Loudoun, Prince William, Fairfax | declared | 0 |
| Central Ohio Fast-growing hyperscale cluster with active county GIS portals. | Franklin, Licking, Delaware | declared | 0 |
| Dallas-Fort Worth, Texas Inside ERCOT, whose generation reporting is machine-readable. | Dallas, Tarrant, Denton, Ellis | declared | 0 |
| Atlanta metro, Georgia Rapid buildout against a constrained transmission network. | Fulton, Douglas, Coweta, DeKalb | declared | 0 |
| Salt Lake Valley, Utah Arid, like the pilot region, which makes water estimates comparable. | Salt Lake, Utah, Tooele | declared | 0 |
| Chicago metro, Illinois Dense interconnection hub; strong county open-data publishing. | Cook, DuPage, Kane, Will | declared | 0 |
| Santa Clara County, California A municipal utility publishes load data no investor-owned utility does. | Santa Clara | declared | 0 |
corporate
- Agency
- Arizona Corporation Commission
- Jurisdiction
- Arizona
Registry note. Would substantially improve shell-company resolution. Held back deliberately pending a privacy review.
- Agency
- US Securities and Exchange Commission
- Jurisdiction
- United States
- Licence
- US Government public domain
- History
- Filings from 2001 onward in full-text search.
Registry note. Useful for confirming operator identity from primary filings rather than inference.
environmental
- Agency
- United States Environmental Protection Agency
- Jurisdiction
- United States
- Licence
- US Government public domain
- History
- Active and historical Clean Air Act permitted facilities.
- Documents held
- 2
Registry note. Connector keeps facilities whose NAICS/name indicate data-processing or hosting and emits backup_generator_air_permit evidence. Ordinary industrial air permits are filtered. Fixtures cover CI when the live API throttles.
Known schema issues
Two-step API: get_facilities returns a QueryID consumed by get_qid. Facility coordinates are sometimes geocoded rather than surveyed, so spatial matches use a distance tolerance. Public endpoints may return HTTP 429 under load.
- Agency
- Maricopa County Air Quality Department
- Jurisdiction
- Maricopa County, Arizona
- History
- Active dust-control registrations.
Registry note. Deferred to a later sprint. Earth-disturbance registrations would be a strong Stage 4 signal if the attribute set can be joined from another AQD service.
Known schema issues
The Dust Control Site layer exposes only ImpactID and geometry; the permit attributes that would make it useful as a construction signal are not published through this service.
infrastructure reference
- Agency
- United States Energy Information Administration
- Jurisdiction
- United States
- Licence
- US Government public domain
- History
- 1990 to the most recent published year.
- Documents held
- 1
Registry note. The supply side of the same question the sales figures answer from the demand side. This is installed capacity, not generation, and it says nothing about whether any particular site could be served -- that would need interconnection data Helios does not have.
Known schema issues
Published only as xlsx. Producer type and fuel source each carry a roll-up row alongside their parts, so reading the sheet naively double counts. Helios keeps only the all-producers, all-fuels row.
- Agency
- United States Energy Information Administration
- Jurisdiction
- United States
- Licence
- US Government public domain
- History
- 1990 to the most recent published year.
- Documents held
- 1
Registry note. The electricity counterpart to the USGS water totals. Its state granularity is coarser than the county water figures, and every row records which it is so the two are never read as comparable.
Known schema issues
Published only as xlsx; EIA offers no CSV equivalent. Provider categories overlap, so only the 'Total Electric Industry' roll-up is read. No public source breaks retail sales to county nationally, so unlike the water totals this cannot be narrowed to a metro area.
- Agency
- US Department of Homeland Security / Geospatial Management Office
- Jurisdiction
- United States
Registry note. This was the intended national upgrade over contributor-dependent OSM coverage. It is declared as withdrawn rather than dropped so the gap stays legible: US substation coverage in Helios is OSM's, with OSM's limits.
- Agency
- OpenStreetMap contributors
- Jurisdiction
- Global (queried for the East Valley study area)
- Licence
- Open Database License (ODbL) 1.0
- History
- Current snapshot only via Overpass. Attic queries could provide history but are expensive and not used.
- Documents held
- 1
Known schema issues
Voltage is a semicolon-delimited string in volts. Completeness is contributor-dependent: absence of a substation is not evidence it does not exist.
land and property
- Agency
- Maricopa County Assessor / Maricopa County GIS
- Jurisdiction
- Maricopa County, Arizona
- Licence
- Maricopa County open GIS data
- History
- Current assessment roll only. Exposes the most recent recorded deed per parcel, not the full transfer chain, which truncates observable ownership history.
- Documents held
- 1
Known schema issues
Date fields are epoch milliseconds. OwnerName is a single unstructured string mixing companies, trusts, and private individuals. Wide LIKE scans are slow (~24s observed), so queries should be bounded by geometry or city first.
- Agency
- Maricopa County Recorder
- Jurisdiction
- Maricopa County, Arizona
- History
- Recorded instruments; deep historical coverage.
Registry note. Helios links out to recorder documents for human verification rather than ingesting them.
municipal planning
- Agency
- City of Mesa
- Jurisdiction
- Mesa, Arizona
- Licence
- City of Mesa open data
- History
- Permits from approximately 2015 onward.
- Documents held
- 1
Registry note. Ingests commercial (COM) permits on East Valley corridors and matches them onto assessor parcels via helios_geospatial.addresses. Residential permits are excluded as noise.
Known schema issues
The public view exposes only permit number, type, address, status, and three dates. There is no valuation, work description, or applicant, and no coordinates, so permits are matched to parcels by normalized address string alone.
- Agency
- City of Mesa Development Services
- Jurisdiction
- Mesa, Arizona
- History
- Various
remote sensing
- Agency
- European Space Agency / Copernicus
- Jurisdiction
- Global
- Licence
- Copernicus open licence
- History
- 2015 onward.
utility and regulatory
- Agency
- Arizona Corporation Commission
- Jurisdiction
- Arizona
- History
- Utility dockets with deep history.
- Documents held
- 2
Registry note. Transmission and substation filings here are the single highest-weight signal in the scoring model, so this gap materially limits Stage 3 recall. It is recorded prominently in the limitations documentation.
- Agency
- Federal Energy Regulatory Commission
- Jurisdiction
- United States — six FERC-jurisdictional RTO/ISO regions
- History
- RM26-4 opened 23 October 2025; six tailored section 206 proceedings opened 18 June 2026.
Registry note. FERC directed each of the six regional operators to justify or reform its tariff through a tailored proceeding. The NYISO order specifically describes a proposed searchable public location for aggregate large-load additions, planned Network Upgrades, and cost estimates. Outcomes and publication formats remain pending, so this entry records a dated coverage gap rather than an ingestible feed.
- Agency
- Michigan Public Service Commission
- Jurisdiction
- Michigan
- History
- Curated official Commission disclosures from December 2025 onward.
- Documents held
- 1
Registry note. The first record is Case U-21990. It reports a contracted load and township, not a street address or parcel, so Helios publishes no geometry.
Known schema issues
Commission news pages are prose rather than a docket API. Discovery is a reviewed URL list, and each new page shape requires a recorded fixture and parser contract before it is published.
- Agency
- Salt River Project
- Jurisdiction
- SRP service territory, Arizona
water
- Agency
- Arizona Department of Water Resources
- Jurisdiction
- Arizona
- History
- Long historical series for groundwater and withdrawals.
Registry note. Needed before any water-use scenario is published. Deferred.
- Agency
- United States Geological Survey
- Jurisdiction
- United States
- Licence
- US Government public domain
- History
- Five-yearly compilations back to 1950. 2015 is the most recent release carrying a county breakdown.
- Documents held
- 1
Registry note. Supplies the reported denominator that Helios's inferred per-site water estimates are shown against. It is a whole-county withdrawal total covering every user, so it is context, never a baseline to subtract from.
Known schema issues
A citation line sits above the real header row. Withdrawals are in million gallons per day; population is in thousands and is rescaled to people on ingest. The 2020 compilation dropped the county breakdown, so 2015 is the newest county figure that exists rather than the newest Helios fetched.